Recent research in individualized (e.g., Lipson, 1967) and computer assisted (e.g., Suppes, 1966) instruction has led to an increasing awareness of the inadequacies of norm referenced testing and the need for testing procedures which determine each individual's mastery on specific types of tasks (e.g., Coulson & Cogswell, 1965). Knowing how well a student has performed relative to some peer group, for example, says relatively little about the kinds of decisions that must be made if instruction is to be totally individualized. Ideally, in mastery testing the procedures used should 1) pravide a sound basis for diagnosing individual strengths and weaknesses on each type of task, 2) require as few items as possible, and 3) provide a basis for generalizing from overall test performance to behavior on a clearly defined universe or domain of tasks.